[ 
https://issues.apache.org/jira/browse/CASSANDRA-1856?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12971101#action_12971101
 ] 

Robert Coli commented on CASSANDRA-1856:
----------------------------------------


Also, the nodetool "help" output is messy :

"
Available commands: ring, info, cleanup, compact, cfstats, snapshot
[snapshotname], clearsnapshot, tpstats, flush, drain, repair,
decommission, move, loadbalance, removetoken, setcachecapacity <keyspace>
<cfname> <keycachecapacity> <rowcachecapacity>, getcompactionthreshold,
setcompactionthreshold [minthreshold] ([maxthreshold])streams
[host],invalidaterowcache <keyspace> <cfname>
"

- should have a ", " between ([maxthreshold]) and "streams"

- and a " " between "[host]," and "invalidaterowcache"

- and [minthreshold] and [maxthreshold] should be <minthreshold> and 
<maxthreshold>, respectively.

> nodetool "snapshot" improvements
> --------------------------------
>
>                 Key: CASSANDRA-1856
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-1856
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Tools
>            Reporter: Robert Coli
>            Assignee: Jon Hermes
>
> Summary : "nodetool snapshot" is actually hooked up to the MBean endpoint 
> takeAllSnapshot. takeAllSnapshot snapshots all CFs in all Keyspaces. 
> It would be good if "nodetool snapshot" took a mandatory Keyspace argument 
> and an optional (list of?) CF name(s?), perhaps mapped to "takeSnapshot" 
> instead of "takeAllSnapshot".
> =Rob
> PS - 
> As a side note, I have never been too pleased with the behavior of "nodetool 
> snapshot" as regards snapshot naming. Most of the time, when someone names 
> something something specific, it ends up actually named that something, but.. 
> "nodetool snapshot foobarbaz" 
> ... does not result in a snapshot at 
> "/path/to/datadir/Keyspace/snapshots/foobarbaz", the snapshot ends up at 
> "/p/t/dd/Keyspace/snapshots/<TIMESTAMPLOOKINGTHING>-foobarbaz".
> When I asked jbellis about this, he basically said it was to avoid having to 
> deal with the possibility that the directory might already exist. If you are 
> mucking about in the snapshot code, you might (or might not!) want to "fix" 
> this behavior. :D

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to